A neuroscientist explains the "Jennifer Aniston neuron," a brain cell that fires only when you see a specific celebrity's face, and how millions of these neurons create rich mental representations. Building virtual worlds could sidestep physical limits, while brain-computer interfaces are already restoring sight to blind people and hinting at stranger possibilities ahead.
I was in 5th grade and literally I saw The Matrix and I was like, I don't know if we're living in a simulation, but we should build one, basically. Look, Earth is small, resources are intensely contested, space is big, the speed of light is low. There are real limits to building with atoms and doing things in the physical world where there's these conflicts, like a piece of land can only be used by one person at a time, building buildings is super expensive and complicated. But like in the simulation, none of those limits are true. In the simulation, there's no scarcity. Have you heard of the Jennifer Aniston neuron? No. So there's a well-known result that there's some, in some part of the brain, if you record from these cells, if you show them pictures of different people, they in some cases will respond very selectively. And so there was a Nature paper a while ago that showed that if you showed pictures of different people, there was a cell that would fire just for pictures of Jennifer Aniston. Or if they showed you types of different houses, they would fire for just specific types of architecture. And so the intuition here is that if you look at all of the millions of neurons in that area, they encode a representation of architecture or of faces or of people or of places. And this is embedded on a mathematical object called a manifold. And so what you want to do is you want to access that statistical representation somehow. And when these neurons grow into the brain and they wire up, you get all this— when the cells in the brain fire, you're getting all this partial information, which represent information about the statistics of those representations. And so you can access this information in and out of the brain through this. This is a very, very deep well of neuroscience that we are, like, at the very beginning of. Just making these devices at all is pretty tricky. I would say this is, like, up there with, like, modern semiconductors for complexity.
They're doing some of the most cutting-edge research and products imaginable in the brain-computer interface field. They're helping people get their vision back, and then they're pushing us possibly in some very weird cyborg-ish directions.
